The explosive growth of telecommunication and computer communications, especially in the area of internet, has placed increasing demand on national and international communication networks. This tremendous amount of worldwide data traffic volume creates a demand for a network having multi-gigabit transmission capacity with highly efficient cross-connect networks. To meet this demand, in the field of fiber optic technology, products have been developed for multi-carrier transmission over a single fiber, which multiplies the amount of capacity over single carrier systems. By assembling several data signals into a multi-channel signal transmitting on a single fiber, commonly referred to as wavelength division-multiplexing ("WDM"), this WDM technology allows multiple users to share a fiber optic link and thereby allowing high throughput over a single link. To assemble the signals, a multiplexing device combines the signals from several sources or channels into a single composite signal. At the receiving end, a demultiplexing device separates the composite wavelength into the several original signals.
There are many specialized applications and variations of the WDM technology. In one variation, dense wavelength division multiplexing ("DWDM") is a WDM device that works for a certain wavelength range and has the ability to handle large number of channels. Some of the critical factors for a DWDM device are high number channels, channel separation, channel spacing, inter-channel cross talk, insertion loss, polarization dependent loss, compactness, environmental stability, and manufacturing cost. To date, wavelength division-multiplexing systems have been unreliable in meeting the aforementioned critical factors,i.e. complete channel separation, low insertion loss, polarization insensitive, etc. Additionally, device size and cost remain high and economically challenging. As a result, there is a need for compact, light weight, low cost devices which can be used as a dense wavelength division multiplexer and demultiplexer compatible with today's fiber optic networks.
There are prior art technologies for making WDM using holographic elements to produce multiplexers and demultiplexers. Despite the fact that prior art technologies provide WDM with highly efficient holographic elements, those devices cannot satisfy the aforementioned critical factors, such as high channel numbers, low cross talk, controlled channel separation and channel spacing. One of the main reasons for the drawbacks in prior art technologies is that, in some cases, there are no optic elements in those devices. Thus it is difficult, if it is not impossible, to correct optical aberrations, such as spherical aberrations, coma, astigmatism, etc. Furthermore, the number of channels provided by these devices are limited because these devices do not provide proper focusing and collimating optics. Rather, they are based on a grating and the arrangement of relative optical fibers with respect to the grating. Also, it is also difficult to reduce cross-talk between adjacent channels due to optical aberrations, which generates beam spots larger than the desirable fiber size, resulting in optical power crossing to other channels in narrow channel spacing situations. For these reasons, these devices are not able to achieve high optical resolution and narrow channel spacing that is critical for DWDM systems.
In another prior art technology, a compact WDM device is provided with spherical grating substrate as a reflective focusing grating. This technology corrects some of the major drawbacks associated with devices without optics. Despite of the fact that the device may provide high spectral resolution, which in turn produce narrow channel spacing, it is not an ideal device. First, the grating is a reflective grating based on surface relief holograms with low diffraction efficiency and thus high insertion loss. Secondly, the grating described in said prior art technology is polarization sensitive, generating intensity variations in the transmitted signals from polarization effects.
Prior art devices having optical elements that are adhered to each other also have high inter-channel cross-talk problems. This is especially the case when glass elements of different refractive indices are used to compensate optical aberrations and the insertion loss from Fresnel (reflection) loss is high. Furthermore, the lost light will be reflected back to the wavelength disperse components either directly or indirectly by the optical element(s) and air. The reflected beam is dispersed by the wavelength disperse components and may be focused, eventually, back to the various components in the system to a position that is different than the expected position. This will result in rather high level of the inter-channel cross-talk. Interestingly, the same drawbacks are also acknowledged in one of the prior art references, U.S. Pat. No. 4,819,224. The required difference in refractive indices to correct the spherical aberration is rather large, causing the chromatic aberration to be rather large and ineffective for industrial applications. Therefore, this type of device cannot satisfy the need for dense wavelength division applications.
Prior art devices are difficult to manufacture. First, some prior art devices involve so-called "locally neutralized" zone that is created on the reflective grating, which is the most critical element of the device. One portion of the grating surface must be treated after the grating is made in order to allow the light to pass. It is difficult to make such a zone without affecting other areas of the grating. Also, this zone causes the grating to lose its integrity and lose significant amount of luminous efficiency. Secondly, some prior art devices involve non-uniform dispersion elements with broad bandwidth, which are not only difficult to manufacture but also may scarify some of the diffraction efficiency. Furthermore, broad bandwidth is practically useless in long-haul telecommunication applications since the erbium-doped self-amplifying optical fibers, which is the enabling technology for wavelength-division and multiplexing in optical networks, only have a limited range of gain spectrum. So broad bandwidth non-uniform grating can only add to manufacturing problems, loss efficiency, resolution and channel separation. Thus, it is not beneficial to DWDM applications. Thirdly, some prior art devices involve off-axis embodiment, which is difficult to align and assemble.
Further drawback of prior art devices is the relative long optical path passing through the many elements. In this type of devices, not only the Fresnel reflection is high. Also, the mechanical and environmental (thermal, stress, etc.) responses of the grating assembly are not monolithic thus not easy to be controlled. They are vulnerable to stress and temperature variations, which may cause inter-channel cross-talk and even variations in channel spacing.
The state-of-the-art wavelength multiplexer/demultiplexer are not suitable for industrial DWDM applications due to the aforementioned major drawbacks, including: manufacturing difficulties, cross-talk, insertion loss, chromatic aberration, luminous efficiency loss, polarization effects, stress and temperature sensitivities, etc. Therefore, there is a need for a DWDM technology to overcome these problems.